Model Filled Rubber. VI: Dynamic Property Dependence on Filler Particle Size of Rubber Compounds During Curing.
The physical and chemical characteristics of fillers present during the vulcanization of rubbery composites influence processability as well as the properties of the cured material. It was reported that, in some filled systems, as the particulate content was increased, the gelation time decreased [1-3]. In other systems, fillers did not alter  or even retarded gelation . To determine optimum processing conditions, it is necessary to understand the chemorheology of thermosetts during curing. This involves the viscoelastic behavior, the variations in viscosity and the growth of the molecular network. In contrast to thermoplastics, relatively little systematic research has been performed on the rheology of filled thermosets during curing. The effect of filler properties such as particle size and chemical composition on the rheology and kinetics of filled polymers during curing has not been elucidated.
A variety of rheological measurements including isothermal or non-isothermal dynamic testing has been developed to assess the curing of thermosets. Hanon  followed curing into the gel state and monitored the building-up of structure via an increase in modulus. Lane  measured changes in the complex viscosity of epoxy systems using an isothermal dynamic time test. Generally, the change from a viscous liquid to a glassy solid is represented by two important viscoelastic transitions: gelation and vitrification [8-10]. On the way to the gel point, the resin viscosity rises rapidly and an irreversible transition from a viscous liquid to an elastic network occurs. As the resin is cured beyond the gel point, the modulus of the network continues to increase until the network vitrifies. The gel point (GP) is defined as the time or temperature at which crosslinks connect throughout the system and a continuous network is formed [11-13]. The gel point is one of the most important kinetic parameters in polymer proce ssing. For example, shaping has to occur before GP, while the polymer is still able to flow and the stress can relax to zero. GP for a crosslinking system has been described [14-17].
G'([omega]) = [[pi]/2[gamma](n) sin (n[pi]/2))] [[S[omega].sup.n] (1)
tan([delta]) = G" ([omega])/G'([omega]) = tan(n[pi]/2) (2)
G(t) = S [t.sup.-n] (3)
t = the present time, 0 [less than] t [less than] s [infinity],
S = the gel strength parameter, depending on the crosslink density and chain flexibility,
n = the relaxation exponent, 0 [less than] n [less than] 1,
S and n are characteristic parameters for the gel structure.
It is inferred that tan([delta]) is independent of frequency at GP. Then, the time of gelation can be measured using a multifrequency experiment in dynamic oscillation rheology. The gel time, [t.sub.c], is obtained as the crossover of a series of curves measured at various frequencies of tan([delta]) v.s curing time, t. Measurements of oscillatory shear moduli are frequently used to monitor the viscoelastic properties of cross-linking systems up to the gel state [18-26]. In this paper, we discuss the rheological properties of filled and unfilled rubber compounds during curing. The effect of particle size on the curing kinetics, as well as dynamic properties will be examined with model filled rubber compounds by using monodisperse size crosslinked polystyrene particles synthesized from emulsifier-free emulsion polymerization.
Monodisperse crosslinked polystyrene (PS) particles with varying particle size were prepared by emulsion polymerization as shown in references [27-34]. The crosslinking agent is divinylbenzene(DVB). Styrene monomer and divinylbenzene crosslinker are products of Aldrich Chemical Company. Styrene is 99% pure and inhibited by 10-15 ppm 4-tertiary-butylcatechol (4-TBC). Divinylbenzene (DVB) consists of a mixture of 55% m- and p-isomers, 42% ethyl vinyl benzene, and 3% diethylbenzene, which is inhibited with ca. 1000 ppm 4-TBC. Styrene and divinylbenzene are stored at 5[degrees]C. The monomer and crosslinker are washed with an equal volume of an aqueous solution of 10% sodium hydroxide for 4 times, followed by deionized water for 4 times, in order to remove the inhibitor before polymerization. Potassium persulfate and sodium hydroxide are certified Fisher Scientific products. The particle diameter ranged from 0.315 to 1.250 [micro]m. The crosslink density was 5 mole% DVB. The curing agent used in this research wa s manganese dioxide [MnO.sub.2] as purchased. Polysulfide with molecular weight Mw [congruent] 8000 g/mole, a viscous liquid with specific gravity of 1.29 g/[cm.sup.3] @ 25[degrees]C, manufactured by Morton International, Inc., was used as matrix. The polymer molecule also contains 0.5 % branched chains. This polysulfide is a polymer of bis-(ethylene oxy)methane containing disulfide linkages. The polymer is terminated with reactive mercaptan(-SH) groups. The molecular structure is as follows:
Filler concentration was 30% by weight for all compounds. Dry powder particles were added to the matrix. All samples were prepared in the same conditions by blending at room temperature with a electrically driven mixer, model Cole-Parmer, series 4401, RPM 60-700, at a rotor speed of 130 rpm for 30 minutes using a 4-blade mixer. Then, the mixing speed raised to 300 rpm for 1 hour. The dense suspension was transferred onto a sheet of polyethylene film, flat-mixed forcefully back and forth using a flat, stainless steel spatula. Finally the blend was mixed at a speed of 300 rpm for 1.5 hours. Samples were stabilized at room temperature for two weeks. The well mixed compound was weighed into a plastic cup and 1.4%w [MnO.sub.2], based on the total weight of polysulfide matrix, was added. The mixture was stirred vigorously for about 4 minutes, then placed onto the rheometer plate and compressed for another 3 minutes before testing. Since curing reaction takes place at room temperature, the time from mixing suspensi on with curing agent to the beginning of rheological measurement was strictly limited to a total of 7 minutes for each sample.
Polysulfide polymer is cured by converting mercaptan(-SH) end groups to disulfide (-S-S-)bonds. This results in a high molecular weight polymer with elastomeric properties. The curing reactions include:
2 RSH + [MnO.sub.2] [right arrow] RSSR + MnO + [H.sub.2]O
2 RSH + MnO [right arrow] RSMnRS + [H.sub.2]O
RSMnSR + [MnO.sub.2] [right arrow] RSSR + 2MnO
Rheological measurements were conducted using a computer-controlled Carri-Med CSL 500 Rheometer in a cone and plate geometry. The bottom plate is a "fixed" surface which is movable (up and down). The top platen is a cone of 4 cm in diameter with a 0.034 radian (2[degrees]) cone angle. Rheological oscillation is ideal for evaluating the development of structure in materials without damaging the structure. The curing process in thermosets can be followed by time sweep of dynamic oscillation measurement performed at a specific frequency during the course of curing reaction. The strain amplitude was set at 2.5 X [10.sup.-3], within which the material was practically in the linear viscoelastic region. The storage modulus (G'), loss modulus (G") were continually monitored as a function of time. All experiments were conducted at 20[degrees]C.
RESULTS AND DISCUSSION
Monodisperse size crosslinked particles, prepared from emulsifier-free emulsion polymerisation, were characterized by scanning electron microscopy (SEM). The particle diameter ranged from 0.315 to 1.250 [micro]m. The crosslink density corresponded to 5 mole% DVB. Particle chemical composition was characterized by FTIR .
Kinetics of Network Formation
Model compounds filled with particles of different particle size were tested in dynamic shear oscillation at three different frequencies. The change in dynamic moduli, G' and G", at a specific frequency and loss tan([delta]) as curing reaction proceeds is shown in Figures 1 and 2 for pure matrix, Figures. 3 and 4 for compounds filled with PS particles of diameter 0.315 micro [micro]m, Figures 5 and 6 for compounds filled with 0.688 [micro]m particles, respectively. The coincidence of tan([delta]) curves obtained at different frequencies permits us to locate gelation: that is, the gelation time is found when tan ([delta]) becomes independent of frequency [14-16]. Since all our rheological measurements started 7 minutes after catalyst was added to the suspensions, 7 minutes must be added to obtain the reaction time. The GP of compounds are found to be 5.5 hours for the pure matrix, 13 minutes for the composite with PS particle diameter 0.315 [micro]m, and 1.4 hours for composite with 0.688 [micro]m PS particle s and greater than 2 hours for composite with 1.25 [micro]m PS particles. The gel time as a function of particle size is summarized in Fig. 7. The gel time decreases as filler particle sizes decrease.
Dynamic Properties During Curing
In this study, dynamic moduli, as a function of curing time, for composites filled with PS particles are compared according to filler particle size. The dynamic moduli, G' and G", are illustrated in Fig. 8 a and b. Both the elastic modulus (G') and the loss modulus (G") increased with decreasing particle size. The compound filled with PS particles, 0.315 [micro]m in diameter, the smallest particle size measured, had the highest moduli during the curing process.
SEM microphotographs of the suspension morphology are shown in Fig. 9 for composite filled with PS/5 mole%DVB particles, diameter 0.315 [micro]m. and in Fig. 10 for composite filled with PS/5 mole%DVB particles, diameter 0.688 [micro]m.
Crosslinking leads to network formation during the curing of thermosets and the vulcanization of rubbers. The storage modulus (G') is related to the elastic behavior of the crosslinking polymer networks, while the loss modulus (G") describes the viscous flow component. The crosslinking of thermosets or rubber compounds during curing can be modelled as a cluster formation process [35-39]. During the initial period of reaction, microgels are first formed in the reacting system before the gel point. Microgels consist of branched and partly crosslinked molecules of colloidal sizes [10, 40]. The polymer chains continue to react and form clusters of various sizes distributed randomly in the composite system. Rheologically, viscous behavior dominates the early stages of curing and G" [greater than] G'. Both dynamic moduli increase as a result of increasing crosslink density and molecular weight in the polymer system. As the curing reaction proceeds, the number and size of clusters or "microgel" particles increase, the molecular weight of matrix polymer increases, and a network forms. The loss modulus (G") continues to increase, but the storage modulus (G') rises more rapidly until it overtakes and exceeds the loss modulus. At gelation, an infinitely large cluster extends throughout the whole system, i.e., a three dimensional continuous network is formed. The storage modulus (G') increases steadily while the loss modulus (G") levels off.
In filled polymer systems, fillers not only influence the curing mechanism but also affect the rheological properties of the composite during curing. All the filled systems showed much shorter gelation times than the pure matrix. That is, the addition of filler accelerated the curing reaction. In an early stage of curing of pure polysulfide matrix (Fig. 1), the material is still liquid and molecular weight of the polymer matrix is low. Then the loss modulus is greater than the storage modulus. However, both moduli are small since branching or clustering is scarce at this time. As the curing reaction progresses, long molecular chains and clusters form and the elastic properties of the system increase. Since intermolecular interactions increase because of increasing molecular weight of the polymer, the loss modulus increases. Although both G' and G" are increasing in this initial stage, G" increases faster than G'. After about 85 minutes of curing, G' increases much faster than G", indicating that a network structure is forming. At about 6 hours, G' surpasses G". The GP for pure matrix as estimated from tan[delta] was around 5.5 hours (Fig. 2).
For compounds filled with PS particles of 0.315 [micro]m in diameter, initially both G' and G" increase, but G' increased much faster. At around the gelation time of 13 minutes, as estimated from tan[delta] (Fig. 4), the storage modulus surpasses the loss modulus, and G' keeps increasing steadily. However, the increase of G" becomes very slow, indicating that an elastic network structure has been formed (Fig. 3). For a compound filled with PS particles, 0.688 [micro]m in diameter, initially, G" increased at a faster rate than G'. However, after 15 minutes, the increase of G" became slower. By 1.36 hours after mixing, G' surpasses G" (Fig. 5). The gel point for this composite is found to be 1.4 hours (Fig. 6).
The dynamic moduli as a function of curing time for composites filled with PS particles of varying particle diameter, from 0.315 to 1.25 [micro]m, are illustrated in Fig. 8 a, b. Both the elastic and the loss moduli increased with curing time for all particle sizes. However. the dynamic moduli increase and the gel times decrease with decreasing particle size. Since PS is not chemically compatible with polysulfide matrix, particle-particle attractive interactions are dominant and lead to the formation of clusters. These clusters or aggregates interact to form a network structure. The clusters act as physical crosslinking points, so that the overall crosslink density of the curing system is increased above that due to chemical crosslinking from vulcanization of the polysulfide matrix. It is shown in Fig. 9 and Fig. 10 that particles of 0.688 [micro]m are more isolated and uniformly dispersed than particles of 0.315 [micro]m in the matrix. We infer, from both rheological behavior and SEM observations, that the ability to form clusters and the effective particle volume fraction increased with increased particle surface area or decreasing particle size . Hence, larger and stronger clusters are formed in compound filled with PS particles of 0.315 [micro]m in diameter compared to those with PS particles of 0.688 or 1.25 [micro]m. Clusters immobilize part of the polymer matrix, effectively increasing the particle volume fraction. The number density, as well as total surface area of particles, increases with decreasing particle diameter. Therefore, as particle size decreases, the colloidal interaction between particles is enhanced. Both physical networks and attractive interactions contribute to the effective crosslinking network so as to accelerate gelation and promote dynamic properties. As a result, the gelation is faster for smaller particle filled composites. The gelation time was reduced from 1.4 hours to 13 minutes as particle size decreased from 0.688 to 0.315 [micro]m in diameter. Under dynamic shear deform ation, the composite filled with smaller particles becomes more difficult to deform, so that more energy will be dissipated. Also with an enhanced network structure, more energy should be stored per cycle of deformation. The dynamic moduli are therefore higher for compounds filled with smaller particles.
The rheological behavior and corsslinking kinetics of model filled composites during curing were investigated by studying dynamic mechanical properties. The effect of particle size of monodisperse particle fillers on chemorheological behavior was studied. Filler particles not only enhance the dynamic moduli of the rubbery composite during curing, but also accelerate the curing reaction. During the curing of filled compounds, both the storage modulus (G') and the loss modulus (G") increase as a function of curing time before the gel point for all systems. However, after gelation, G' keeps increasing steadily while G" changes very slowly. In the process of curing, the magnitude of the dynamic moduli increased and the gelation time (GP) decreased as PS particle diameter decreased from 1.25 to 0.315 [micro]m. GP was 5.5 hours for the pure matrix, 13 minutes for the composite containing 0.315 [micro]m particles, 1.4 hours for 0.688 [micro]m particles and more than 2 hours for 1.25 [micro]m particles.
In the model PS particle filled systems, apparently, the ability to form clusters was due to interparticle dispersive interactions which increase with decreasing particle size. The particle clusters function as physical crosslinks. As a result, the overall crosslink density (chemical plus physical) was significantly enhanced. Both the dynamic moduli and the rate of gelation were increased with decreasing particle diameter.
The authors are grateful to the Los Angeles Rubber Group, Inc. (TLARGI) for the financial support in this work.
(1.) H. Kubota, J. AppL Polym Sci., 19, 2279 (1979).
(2.) C. D. Han and K. W. Lem, J. AppL Polym. Sci., 26, 3185 (1983).
(3.) H. Ng and Ica Manas-Zloczower, Polym. Eng. Sci., 33(4) (1993).
(4.) Weiling Peng and Bernard Riedl, Polymer, 35 6), 1994).
(5.) M. Paauw and A. Pizzi, J. Appl. Polym. Sci., 50, 1287 (1993).
(6.) M. J. Hannon, D. Rhum and K. F. Wissbrun, J. Coat mgs Technol., 48, 42 (1976).
(7.) J. W. Lane, J. C. Seferis, and M. A. Bachman, Polym. Eng. Sci., 26, 346 (1986).
(8.) J. B. Enns and J. K Giliham, J. Appl Polym. Sci., 28, 2567 (1983).
(9.) J K. Gillham, Brit. Polym. J., 17, 224 (1985).
(10.) A. Ya Malkin and S. G. Kulichikhin, Advances in Polymer Science, 101, 218 (1991).
(11.) P. J. Halley and M. E. Mackay, Polym. Eng. Sci., 36, 593 (1996).
(12.) C. W. Macosko, Brit. Polym J., 17. 239 (1985).
(13.) B. Erman, Crosslinking and Scission in Polymers, 292, 153, Nato ASI Series, Kluwer Academic Publishers, Dordrecht, The Netherlands (1988).
(14.) H. Winter, F. Chambon, J. Rheol., 30, 367(1986).
(15.) H. H. Winter, Polym. Eng. Sci., 27, 1698 (1987).
(16.) F. Chambon, H. H. Winter, J. Rheology, 31, 8. 683 (1987).
(17.) H. H. Winter, Progr. Colloid Polym. Sci., 75 (104) (1987).
(18.) E. M. Valles, J. M. Carella, H. H. Winter, and M. Baumgaertel, Rheol. Acta, 29, 535 (1990).
(19.) J. C. Scanlan and H. H. Winter, Macromolecules, 24,47 (1991).
(20.) A. Izuka, H. H. Winter and T. Hashmoto, Macromolecules, 25, 2422 (1992).
(21.) M. Adam, M. Delsanti and D. Durand, Macromolecules, 18, 2285 (1985).
(22.) D. Lairez, M. Adam, J. R. Emery, and D. Durand, Macromolecules, 25, 286 (1992).
(23.) D. F. Hodgson and E. J. Antis, Macromolecules, 23, 2512 (1990].
(24.) R. Muller, E. Gerard, P. Rempp and Y. Gnanou, Macromolecules, 24, 1321 (1991).
(25.) B. S. Chiou, R. J. English, S. A. Khan, Macromolecules, 29, 5368 (1996).
(26.) A-L. Kjoniksen, B. Nystrom, Macromolecules, 29, 5215 (1996).
(27.) J. F. Cai, PhD dissertation, University of Southern California (1997).
(28.) D. Zou et al., J. Polym. Sot: Part A: Polym. Cherm., 28, 1909 (1990).
(29.) D. Zou, S. Ma, R. Guan, M. Park, L. Sun, J. J. Aklonis, and R Salovey, J. Polym. Sot: Part A: Polym. Chem., 30, 137(1992).
(30.) Z. Y. Ding, S. Ma, D. Kriz, J. J. Aklonis, and R. Salovey, J. Polym. Sci.: Part B: Polym. Phys., 30, 1189 (1992).
(31.) D. Zou, L. Sun, J. J. Aklonis, and R. Salovey, J. Polym. Sci.: Part A: Polym Chem, 30, 1463(1992).
(32.) J. Lee and M. Senna, Colloid Polym. Sci., 273, 76 (1995).
(33.) M. Okubo and T. Nakagawa, Colloid Polym. Sci., 272, 530 (1994).
(34.) L. Sun, PhD dissertation, University of Southern California (1992).
(35.) M. Takahashi, K, Yakoyama, and T. Masuda, J. Client Phys., 101 (1), 798 (1994).
(36.) Hess Walter, Thomas A. Vilgis, and H. H. Winter, Macromolecules, 21, 2536 (1988).
(37.) T. A. Vilgis, Progr. Coilold Polym. Sci., 90, 1 (1992).
(38.) M. Muthukumar, Macromolecules, 22,4656(1989).
(39.) B. Sun and T. L. Yu, J. Appl. Polym. Sci., 57, 7(1995).
(40.) B. E. Moore, L. S. Lee, J. Appl. Polym. Sci., 8, 625 (1964).
(41.) J. F. Cal and R. Salovey. Polym. Eng. Sci., submitted (1998).
|Printer friendly Cite/link Email Feedback|
|Author:||CAI, JIANFEN J.; SALOVEY, RONALD|
|Publication:||Polymer Engineering and Science|
|Date:||Jan 1, 2000|
|Previous Article:||Film Casting of Polyethylene Terephthalate: Experiments and Model Comparisons.|
|Next Article:||Prediction of the Gas Permeability of Heterogeneous Polymer Blends.|